Embracing AI-Driven Decision Making in Manufacturing

AI-driven decision making in manufacturing delivers measurable results when leadership commitment, operator trust, and incremental scaling align. In a conversation between Alec Glenn of JSW Steel USA and Karthikeyan Natarajan, Co-CEO of Infinite Uptime, concrete outcomes emerged: one tire plant increased output by 15% in a year without adding equipment, another facility achieved a 2.5% capacity improvement, and a third site saw Mean Time Between Failures increase by 70%, all from acting on prescriptive AI recommendations.

It Doesn’t Begin with Technology

JSW Steel’s journey didn’t start with sensors or models. According to Alec, it started with leadership. The Jindal family and JSW’s executive team — in India and the U.S. — agreed on the direction and committed resources. Only after that did AI tools make their way to the shop floor.

This mattered because, as Alec pointed out, it’s unrealistic to expect technicians to trust a new system if leadership hasn’t shown that it matters. And despite all the conversations around frontline “buy-in,” he was honest: they’re still working toward that. The “aha moment” isn’t a single event — it’s something still forming.

What AI Has Actually Delivered Elsewhere

To give the discussion some grounding, Karthikeyan Natarajan — Co-CEO of Infinite Uptime — shared results he has seen from manufacturers already using prescriptive maintenance and AI tools in live production environments. One tire plant increased output by 15% in a year without adding new equipment. Another facility recorded a 2.5% capacity improvement. At a different site, Mean Time Between Failures went up by 70%. These weren’t projections or marketing numbers — they came from operating teams acting on AI-driven recommendations. 

These numbers came from technicians and engineers who chose to use AI recommendations because they started to see them working.

Trust Takes Time, Especially on the Plant Floor

Trust doesn’t come from presentations; it comes from repetition. Alec explained that maintenance teams at JSW Steel began to take AI seriously only after it consistently spotted issues early, communicated clearly, and didn’t disrupt their work. He stressed something simple but rarely acknowledged: success has to be proven quietly, not declared loudly.

And when AI gets it wrong? Workers sometimes ignore it. That’s not seen as failure — it’s part of the feedback loop. The system adjusts, and people stay involved. Alec emphasized that forcing AI onto people without room for human judgment is a fast way to lose them.

Starting Small Instead of Rolling It Out Everywhere

Instead of launching AI across every site at once, JSW chose a slower route. Plants in India are further ahead. In the U.S., they’re starting with specific assets — pumps, hydraulic systems, gearboxes. The idea is to get it right in one area, let people see the results, and then carry that experience to the next plant.

This isn’t scaling software. It’s scaling proof.

This article is based on the insights shared by Alec Glenn of JSW Steel USA and Karthikeyan Natarajan of Infinite Uptime during an interview with Lucian Fogoros of IIoT World. 

Sponsored by InfiniteUptime